Roy10266
How are you unable to say for sure? Do you guys not have a dozens of examples, or do forum guys not get any actual info about what matters most? I'd totally be fine to know that the FTW DT I ordered by mistake could only reach 2050 MHz core instead of the regular 2075-2100. That's only 1 or 2 frames difference. But if I can only go 2GHz, when the FTW 1070 can easily do 2075, I will be so disappointed.

I mean, I could return it before opening it. But I wonder how much that'll end up costing myself in shipping.

Scarlet-Tech
Brimy, it means down tuned. It runs at Nvidia stock clocks.

pharmacist329
I unknowingly bought the DT thinking it was the FTW based on previous branding. I would have been concerned, but I know from experience that many of these cards are paced out on the pricing scheme due to tiny variations in performance. Most of these tiny variations are attributed to slight bios overclocks such as allowing for more voltage. Basically I went in knowing that I was going to modify the bios no matter what, so I just wanted the best value card, phases etc. So, I get the card, plug it in, turn on kboost via evga's software, and the card boosts to what it says sure, but then I overclock it to 2050mhz. It has passed every benchmark I have thrown at it so far at 2050mhz with a +415 mem offset. Keep in mind, the voltage has never really reached 1090, and I monitor it while I play and bench. I used to run my world record 970s are 1.28volts on air. This means there is a ton of room for voltage, especially since when I am gaming I am really only seeing like 65c max on 60% fans, usually hovering around 60c. So, although I am sure evga has run a gauntlet of tests, I can't say that it truly amounts to much in the end, whether I bench at 2000mhz or 2050mhz, the scores are fairly identical. If you consider yourself a real overclocker, get in the bios and up the voltage limit with the pascal tweaker comes out, thats where the real value is; not in spending 679, 700, 710, 750 dollars on something that gets 3 better frames per second on similar gimped bios. For what its worth, I have seen a lot of youtubers having trouble hitting 2100 in all applications, founders or not, now consider that may be a voltage problem. If you open gpu-z and turn your eyes to the sensors you will see that the performance is voltage capped, meaning if there was more volts fed to the card it would perform better, it is not yet hitting its potential. For instance, when I upped the voltage on my 970s to 1.28, with kboost they automatically clocked at 1500mhz, and from there I overclcocked about 70 more. My point is, if you got a FTW DT, if you are eyeing one in stock right now and considering if you should pass it up; from my personal experience, I don't think you should. These cards are hard to come by in stock, and you NEED that evga software to get good overclocking and OSD elements, any other program is tedious and lacking. 
 
TLDR Get the FTW DT, OC to 2050, be happy with 200 fps instead of 200.13
 

ganzosrevenge
I would be 100% fine with a DT version if the box said on the front "REF" to inform the customer that this card has the structure of an FTW but the speeds of the reference cards.  The 980 Ti Classy with the Ref. Clock speeds said "REF" on it to denote this.  The problem I have is calling something DT.  DT is not something that would be commonly known to someone... does it mean "double tuck?" or "detuned" or "down there?"  If that was made known in plain terms to the consuming public, I would have no problems with a Reference FTW for those who do not want to OC the GPU beyond what it can do, but still want the benefits of smoother power delivery, RGB lighting, and a far superior air-cooler to what's on the reference design.